A DNA Sequencer Cheap Enough For (Some) Doctors' Offices

cylonlover writes "Until recently, DNA decoding machines — fitting in the US$500,000 to $750,000 price range — would take weeks or even months to sequence a human genome, and the whole procedure would cost $5,000 to $10,000. That could be about to change, however, as Life Technologies introduces the Benchtop Ion Proton Sequencer — a machine that may finally deliver the power of genetics into the hands of ordinary doctors thanks to its $149,000 price tag and ability to decode a human genome in one day at a cost of $1,000."

Comments

Unforeseen consequences

[bonch]

Warning--side effects may occur.

Re:Unforeseen consequences

[grantek]

This is a completely different machine, it's not the one that TSA agents will use to remotely scan your DNA...

Re:Unforeseen consequences

[justforgetme]

omg, the days of GATTACA are finally here! Now I'll never fly to mars ${sadface}

Re:Unforeseen consequences

[zippthorne]

Half of what was happening in Gattaca. Not even the interesting half.

Messing with genes was the logical response to the first half, though, which was gene discrimination. Or did you not notice that the main character was smart enough and driven enough to become a rocket engineer, but because of a chance of a flaw in other areas was relegated to menial labor.

It's not just that they wouldn't let him be an astronaut, either. They wouldn't even hire him as an engineer at all, as a "bad risk." And his love-life was implied to have suffered as well, with the matchmaking sequencers on every street corner....

The movie was about the horrors of discrimination, and the virtue of overcoming them, not gene manipulation, which is not a necessary precondition to gene discrimination. Regardless, I think it was probably supposed to be an allegory to present-day race discrimination, but with a narrative trick to make the character white, so white people wouldn't have any preconceived notions getting in the way of the message, rather than a direct prediction of the future, however prescient it may appear to have been.

Re:Unforeseen consequences

[cbiltcliffe]

They also conveniently forget to mention the "side effect" that you need a $75,000 computer setup to analyze the data the $149,000 machine spits out. So in reality, this is actually a $225,000 machine, plus an IT person to manage the hardware.

Re:Unforeseen consequences

[mrops]

Gene discrimination has been happening for ever, some call it Evolution.

Hospitals

[Anonymous Coward]

I predict that the first buyers will be University research hospitals, and the Mayo Clinic.

It needs to drop a bit more before seeing it at your local pediatrician's.

Re:Hospitals

[Samantha Wright]

Gene hacking already is the next nerd occupation (I should know; I'm in the middle of it. Mozilla even funds projects for it.) Here's one starting place if you're really interested.

About the read/write thing: synthesizing large amounts of DNA from scratch still costs ungodly amounts of money. Further, the ABI IonTorrent system being advertised here is a destructive read; you have to treat a blood sample with a large number of chemicals and then stuff it in a big machine. It's no Star Trek scanner.

Re:Hospitals

[Samantha Wright]

Yes, but it's fraught with errors, and making a single molecule longer than a few thousand bases costs a great deal. A bacterial genome is 0.4-3 million bases; humans are 3.1 billion in total. That's why only the Venter Institute has done it.

Re:Hospitals

[bill_mcgonigle]

It needs to drop a bit more before seeing it at your local pediatrician's.

And I'm not sure that it ever will. How long has generating a CBC been a solved science? But the pediatrician will still send you over to the hospital's phlebotomy lab to get one.

I guess they do do instant strep tests in the office. When a DNA sequence costs $15 including equipment perhaps they'll do it.

Re:Hospitals

[cperciva]

The only reason they do instant strep tests at MD offices is that they want to know the results before you walk out the door. The same goes for rapid STD tests: If you walk out the door and have to come back to get the results later, there's a chance you won't return.

A DNA test is the ultimate in non-urgent tests, and is going to remain something you head off to the phlebotomist to get done... right up until the day where they are done routinely at birth and you leave the hospital jugging a baby, birth certificate, and a flash drive containing its DNA sequence.

Re:Hospitals

[RDW]

... right up until the day where they are done routinely at birth and you leave the hospital jugging a baby, birth certificate, and a flash drive containing its DNA sequence.

Which might happen sooner rather than later:

http://www.forbes.com/sites/matthewherper/2012/01/10/not-quite-the-1000-genome-but-maybe-close-enough/

"Yale geneticist Richard Lifton, who was the first to document the use a DNA sequence to diagnose a disease, is looking at utilising the Proton for clinical work. In the state of Connecticut, where Yale is based, infants are tested for 43 different genetic mutations that need to be detected early in infancy. The Proton could be a better way to do that that traditional methods, especially given its ability to deliver results quickly."

There's going to come a point when even whole genomes are cheaper to do than a few dozen separate conventional tests.

Don't believe the hype

[Anonymous Coward]

There are two unfortunate challenges that the Ion Proton approach hasn't yet solved. The first is that the steps required to get the DNA out of human cells and into the sequencer (DNA extraction and especially library preparation) are still frustratingly complex. Their OneTouch device simplifies parts of the library prep but there are still many steps that require highly skilled people doing hours to days of work.
The second major issue is that the genome is being read out in fragments of 200-400 nucleotides, then needs to be assembled. The human genome is full of repetitive regions that are much longer than 200-400nt and when one gets a sequence read from one of these regions, it's can be very difficult to determine which of the copies of the repeat region that sequence came from. Better statistical models and algorithms for genome assembly may solve this to some extent, but there are fundamental limits to what can be done with short sequence reads. Other sequencing technologies don't suffer the short read problem, Pacific Biosciences' hardware for example can read several thousand nucleotide fragments. Mate pairing strategies might be used on the Ion instrument but the library prep for these involves considerably more challenging and manual lab work.

Re:Don't believe the hype

[ColdWetDog]

The third problem (once you solve the first two) is what the hell you do with all that information. At present, having a complete readout of your genome doesn't get you very far. Even after you've figured out what diseases that you are more at risk for, what do you do? Well, you keep an active, healthy lifestyle, drink in moderation, don't smoke etc. You didn't need all of that info for me to tell you that.

I don't see this in doctor's offices (except for the boutique practices that do everything to / for you for the specific purposes of lightening your wallet). Maybe it will allow smaller research groups to tackle projects that they couldn't afford to do. But it's a long way to clinical utility.

Re:Don't believe the hype

[Samantha Wright]

Pacbio promises a trillion unicorn farts per second. It's hard to take them seriously. As far as medical applications are concerned, the read length in the Ion Torrent system is ten times the size it needs to be, since most (known) diseases occur due to mutations in the very specific and non-repetitive exome, or in its close vicinity. No one (that I know of) has ever seriously proposed using this hardware for de novo sequencing of large eukaryotes, especially since the machine currently on offer doesn't have the well capacity to sequence the whole human genome!

That being said, there's always paired-end reads. I'm guessing the protocol for doing so with this system doesn't exist yet, but they tend to solve most of the repetitiveness problems for shorter read lengths.

Re:Don't believe the hype

[Samantha Wright]

You archive it until you need it. A situation-specific microarray might cost a hundred dollars; those tend to stack up with every hospital visit. With whole exome sequencing like this, you pay the fee once, and have all* the data medical science will ever need about you.

* Not counting repetitive elements, promoter regions, UTRs, spacer DNA, or the epigenome, all of which are known to describe at least a few diseases.

Re:Don't believe the hype

[backslashdot]

"the read length in the Ion Torrent system is ten times the size it needs to be"

Uh no. Not at all.

"most (known) diseases occur due to mutations in the very specific and non-repetitive exome"

The problem is that if it does occur outside of this, there's be no way to tell where exactly it is. Second, even if it is within the non-repetitive region .. a mutation could make matching the sequences difficult since it wouldn't be certain if you're dealing with an overlap or a mutation. The reason most known diseases occur due to mutations in non repetitive areas is because those are the easiest areas to detect. The unknown diseases probably occupy the other spots.

For disease specific mutations such as cancer the only way to detect the mutations is with long reads. If you want to cure cancer there needs to be a way to do long reads with single cell sequencing.

Re:Don't believe the hype

[Anonymous Coward]

The third problem (once you solve the first two) is what the hell you do with all that information. But it's a long way to clinical utility.

Well, if by "long way" you mean five years, then yes, it's a long way.

On the other hand, in five year when this technology gets of the ground, it will transform diagnostic medicine in the same way that the automobile transformed personal transportation.

One huge application is diagnosis of infectious disease. Not only will you know whether what you have is viral or bacterial, if it's bacterial you'll know exactly which antibiotics will work. And you'll probably even be able to correlate you infection with known outbreaks: "There's been 50 other infections with this pathogens in the northeast corner of this city over the last two weeks."

The other huge application is birth defects. Many birth defects are recessive and this technology will tell couple which birth defects they are both carriers for - and could pass on to their children. Also, in the case of most mental retardation, the actual causative mutation isn't known (it doesn't have distinct symptoms) but when the comprehensive databases of human mutations start coming online, it will be possible to know exactly which mutation is causing the problems. And, in rare cases, that will even lead to successful therapy (the other cases will have to wait for general gene therapy).

Bottom line: high throughput sequencing for medical diagnosis is going to be one of major advances of this century.

Re:Don't believe the hype

[lbbros]

There's another problem to bear in mind: coverage. What is coverage exactly? In terms of "next generation sequencing" (to which this machine belongs to) is how well a part of the genome is covered (sequenced), and that in turns means that the number of fragments (100-150 in the case of Ion Torrent) read for a specific region must be high as possible (those are called "reads").

A good coverage allows you ensure that what you're seeing is real and not some sequencing errors (all technologies suffer from certain types of errors). In the case of Ion Torrent, you usually sequence fragments up to 1 Gbp of sequenced material (G base pairs of DNA), which isn't enough to cover the inter-genic regions or other structural DNA. So usually you do a "reduction step" by either only selecting a handful of genes to be sequenced (mutation screening for example) or by sequencing only the "exome" (exons are the parts of DNA that are actually transcribed to mRNA, the exome is a collective term for all the exons in the genome). That may be useful per se but it's nowhere near a complete genetic map for one individual.

Equipment to do that is much more expensive (500K USD, not counting reagents and other machines).

Re:Don't believe the hype

[RDW]

With whole exome sequencing like this, you pay the fee once, and have all* the data medical science will ever need about you...* Not counting repetitive elements, promoter regions, UTRs, spacer DNA, or the epigenome

Not just exome, but whole genome in about a year (they claim), so everything except the epigenome. Some interesting discussion here:

http://pathogenomics.bham.ac.uk/blog/2012/01/ion-torrent-proton-the-chip-is-not-the-machine/

http://seqanswers.com/forums/showthread.php?t=16709

Re:Don't believe the hype

[RDW]

For disease specific mutations such as cancer the only way to detect the mutations is with long reads.

This really isn't true. Though it would be nicer to have longer reads, short read sequencing (exomes or whole genomes) is actually proving to be very successful at picking up many novel mutations in cancer. Discovering tumour-specific somatic mutations is a particularly nice application for this technology, as you can often do a direct pairwise comparison with the patient's normal germline DNA (e.g. blood DNA if you're working with solid tumours). Typically results will be validated by another technique (e.g. conventional Sanger sequencing), but with a decent analysis pipeline you shouldn't get too many false positives from the NGS.

Re:Don't believe the hype

[Rutulian]

One huge application is diagnosis of infectious disease. Not only will you know whether what you have is viral or bacterial, if it's bacterial you'll know exactly which antibiotics will work. And you'll probably even be able to correlate you infection with known outbreaks: "There's been 50 other infections with this pathogens in the northeast corner of this city over the last two weeks."

I think you're overselling this a bit. Sequencing a human genome is not going to tell you whether you are infected with a pathogen, and it won't identify the pathogen. To do that, you will need to isolate the organism and sequence it that way. Also, having the genome of an organism does not automatically make it easier to determine things like antibiotic resistance. In some ways it makes it more difficult, because you have too much information. High throughput sequencing is definitely a great advance for the scientific community but, like all great advances, it's a tool that is useful for some things, not a magic answer machine.

Oh and by the way, everything you mention for diagnosis above can already be done with current methods faster and cheaper than with whole genome sequencing.

How does that compare to X-ray machines?

[jcr]

A hundred and fifty grand doesn't sound like a whole lot of money for medical equipment.

-jcr

Re:How does that compare to X-ray machines?

[ColdWetDog]

"Ah, I see you have the machine that goes ping. This is my favorite. You see we lease it back from the company we sold it to and that way it comes under the monthly current budget and not the capital account. "

Yay, progress!

[martin-boundary]

Patient: "So, snif, bhat dh'you bheckon I habh, snif, bhoctor?"

Doctor: "Well, it looks like you have a common cold. But let's be sure, shall we? I just got this new DNA sequencing machine. Come back tomorrow."

The next day...

Patient: "Hebho bhoctor, bhat dho I habh?"

Doctor: "Well, it looks like you have a common cold. That will be $1000."

Re:Yay, progress!

[LordKronos]

Huh? In what way does the cold virus alter your DNA, such that a DNA sequencing machine could be used to diagnose it? And you know, doctors currently have all sorts of diagnostic tests at hand that could be used unnecessarily, yet I've never heard of doctors using an x-ray, EKG, MRI, or anything of the sort to diagnose a cold.

Yeah, I know, it was probably a joke and not meant to be taken seriously, but your score was +3 Insightful, not +3 Funny.

Re:Yay, progress!

[cbiltcliffe]

Doctor: Hmm, you're still early in the infection so it could just be a virus but let's use our high throughput sequencer to check. Just cough into this tube.

  Doctor (a couple hour later): Yes, you do actually have walking pneumonia

You know those crime shows on TV that where the cop and a computer guy have a crap video from a 320x240 surveillance camera, and the cop asks the computer guy "Can you enhance that?" and the video zooms right in on the suspects face, which then turns from about a 9 pixel smudge into a 10 Megapixel image that's clear enough to see the pimple on the perp's nose?

Yeah....

Just so you know, high throughput sequencing is currently available, and has been for a year or two. The problem is all the lab work that needs to be done to isolate the DNA, purify it, and put it into a solution that this instrument can analyze. (Not to mention in a case like this, the viral DNA would also have to be separated from the patient's DNA, which would probably at least double the processing time.)

All that stuff is manual lab work that can not currently, or for the foreseeable future, be automated, and it takes much longer than a couple of hours.

Hands of ordinary doctors

[nbauman]

Yes, you bring your doctor a thumb drive with 3 billion base pairs of your genome, coding for 23,000 genes. Do you know what he says?

"What am I supposed to do with that?"

Years ago, people thought that we could find Mendelian genes for all the important things in health and disease. Now it turns out that most of the important things we want to know are controlled by hundreds or thousands of genes, each of which increases the risk by 1%, sometimes less. That's for things like cholesterol, autoimmune diseases, cancer susceptibility, etc.

For the most part, your family history is a better predictor than any genome screening. Gene tests usually aren't useful unless you have a particular gene in your family and you want to find out whether you have it, like the BRCA genes for breast cancer. If your mother died of breast cancer at age 40 because of the BRCA1 gene, and you don't have the BRCA1 gene, you don't have to worry.

Re:Hands of ordinary doctors

[pesho]

Yes, you bring your doctor a thumb drive with 3 billion base pairs of your genome, coding for 23,000 genes. Do you know what he says?

"What am I supposed to do with that?"

You hit the nail on the head. There is very little if any useful information for a doctor in full genome sequence, and most of it can be obtained with much cheaper genetic tests. ABI is ramping up the hype because they really need this instrument to be a good seller. Their first bet in the field was on a sequencing by ligation machine (SOLID) which did not sell very well.

Re:What does it all mean?

[Samantha Wright]

There are a few key puzzles we need to crack—protein dynamics, mostly—and then we'll have the ultimate acceleration method: we'll be able to simulate it all. Right now we're content with trying to sort out those challenges, and fixing the disorders we already know how to recognize. It'll keep us busy for a long time. In the interim we'll just play with increasingly clever tricks to sort out the patterns in the sequences, and catalogue lots of people so we have things to work with when the time comes.

Re:Another medical money-grubbing bullshit

[Samantha Wright]

The term "junk DNA" is now only used by shoddy science journalism. We're quite comfortable with how DNA and RNA do what they do. There's a mystery about what happens on the protein side, and the question about what functional bits of RNA (called microRNA) interact with what genes is sheerly a matter of ridiculously obtuse combinatorics. Say whatever else you will about them, fat cancer research budgets have taught us a lot about the essentials.

Re: Your sig

[srussia]

I am a biologist. Ask me questions in my journal. I'll give car/computer analogies if possible!

No need for the invite. This is Slashdot. You had us at "Samantha".